Face recognition using modified 3D mixed techniques

العناوين الأخرى

تمييز الوجه باستخدام تقنيات مدمجة مطورة ثلاثية الأبعاد

مقدم أطروحة جامعية

al-Tai, Hamid Musa Hasan

مشرف أطروحة جامعية

al-Jawhar, Walid Amin Mahmud
Ulwan, Majid Abd al-Nabi

أعضاء اللجنة

Ali, Ramzi Salim
Ali, Hamzah A.
Ali, Abd al-DhIm A.
Majid, Ghassan H.
al-Saffar, Ala Abd al-Wahhab Hasan

الجامعة

جامعة البصرة

الكلية

كلية الهندسة

القسم الأكاديمي

قسم الهندسة الكهربائية

دولة الجامعة

العراق

الدرجة العلمية

دكتوراه

تاريخ الدرجة العلمية

2012

الملخص الإنجليزي

Face recognition (FR) is one of the most important biometrics which seems to be a good compromise between actuality and social reception and balances security and privacy well.

It has a variety of potential applications in information security, law enforcement, access controls and so on.

Face recognition systems fall into two categories : verification and identification, Face verification is 1 : 1 match that compares a face images against a template face image.

On the other hand face identification is1 : N problem that compares a probe face image against all image templates in a face database.

Face recognition is a very difficult problem due to a substantial variation in light direction (illumination), different face poses, and diversified facial expressions, Aging (changing the face over time), Occlusions (like glasses, hair, cosmetics).

In this thesis two methods for face recognition were developed.

Two dimensional face recognition (2D FR) was employed in the first method.

In this method a modified Radon transform was used for features extraction.

A subset of the features extracted is selected using the particle swarm optimization (PSO).

The selected features are further reduced into a smaller set of features by using the Principal Components Analysis (PCA).

The classification is carried out using the Linear Discriminant Analysis (LDA).

This method was applied on ORL face database.

Five experiments were conducted.

A recognition rate of 97.5 % was achieved.

On the other hand, three dimensional face recognition (3D FR) was employed in the second method.

In this method a modified 3D Radon Transform was used for the feature extraction.

A small subset of the huge extracted features was selected using the particle swarm optimization (PSO).The Discrete Wavelet Transform (DWT) was used in order to reduce the selected features into one half, that is by filtering out the detailed part of DWT output, and using only the approximation part.

This method was applied on the Texas 3DFR database.

Six identification experiments and one verification experiment were conducted.

In the identification experiments a closed universe with disjoint data sets was used.

Different classification methods were used.

One of these methods consists of using a linear discrimination functions, derived by the LDA, to represent each class (subject).

Then the matrix of the class probability was used for classification.

This was applied on 18 subjects with 10 images per subject, and a recognition rate of 100 % was achieved.

The recognition rates achieved by the other experiments were within (81 %, 82 %, 90.9 %, 96.6 %, and 99.44 %).

For the verification experiment an Equal Error Rate (EER) achieved was 1.8 %.

In this experiment an open universe with true imposters was used.

التخصصات الرئيسية

الهندسة الكهربائية

الموضوعات

• معالجة البيانات

عدد الصفحات

136

قائمة المحتويات

• APA
• MLA
• AMA

لغة النص

الإنجليزية

نوع البيانات

رسائل جامعية

رقم السجل

BIM-317374